Bounti — agentic threat model
Bounti acts as an AI sales teammate that automates prospect research, pitch generation, and multichannel outreach. Its primary security risks stem from potential prompt injection via untrusted external prospect data (e.g., LinkedIn profiles) and the risk of automated brand damage or phishing if outreach tools are compromised.
OWASP AIVSS score rationale
| Autonomy of Action | 0.50 | |
| Goal-Driven Planning | 0.40 | |
| Self-Modification | 0.10 | |
| Dynamic Tool Use | 0.50 | |
| Persistent Memory | 0.40 | |
| Contextual Awareness | 0.70 | |
| Dynamic Identity | 0.20 | |
| Multi-Agent Interactions | 0.10 | |
| Non-Determinism | 0.60 | |
| Opacity & Reflexivity | 0.50 |
Scored with the canonical OWASP AIVSS formula (AIVSS calculator reference); agentic risk factors estimated from the agent’s described capabilities.
MAESTRO 7-layer threat model
Per-layer threats for this agent. Layers tagged “not certain from listing” are general, caveated commentary where the public description didn’t pin that layer.
Not certain from the listing — likely utilizes commercial LLMs for generating personalized pitches and role insights. It is highly vulnerable to indirect prompt injection if malicious instructions are embedded in the prospect websites or LinkedIn profiles it scrapes.
Not certain from the listing — ingests external company and role data to perform RAG. This introduces risks of data poisoning and integrity issues if the source data (public web pages, social profiles) is manipulated by external actors.
Not certain from the listing — orchestrates workflows to generate battlecards, landing pages, and outreach messages. Insecure tool integration is a risk if the framework executes API calls to email or LinkedIn networks without strict schema validation.
Not certain from the listing — hosted as a closed-source SaaS platform. Key threats include the exposure of sensitive API keys and OAuth tokens used to connect to CRM systems, email servers, and social media accounts.
Not certain from the listing — no guardrails or output monitoring systems are detailed. Without robust observability, the agent could generate and distribute hallucinated, inappropriate, or brand-damaging content to prospects.
Not certain from the listing — being a closed-source freemium tool, there is no explicit mention of compliance certifications (e.g., SOC2, GDPR). Handling prospect PII and corporate sales data without clear compliance controls poses regulatory risks.
Not certain from the listing — functions primarily as an isolated AI teammate but integrates deeply with external communication ecosystems (Email, LinkedIn). A compromise could lead to downstream trust abuse, where recipient systems treat Bounti's automated spam or phishing as legitimate user activity.
MAESTRO — the 7-layer agentic threat-modeling framework (Cloud Security Alliance / Ken Huang).